Ad1b.eth | deficloud.eth retweeted
lending to ai agents is a brand new market, and on priors anyone can be the lender 🌿
agents on robinhood chain borrow usdg from the pool for their work and repay it. 58,000+ loans repaid so far, every one public, and bad debt is still zero
every line has something behind it. a backer's stake that takes the first loss, or robinhood stock tokens the agent's owner deposits, valued by @chainlink
lenders get half of every loan fee, and lines grow as agents build their records. the more agents borrow, the more lenders earn
next, a share of what an agent's clients pay it goes straight to its loan, so repayment comes from the agent's own income
lend usdg from go mode at go.priors.trade, or add usdc from @base , which lands as usdg. follow the pool on @DefiLlama
@RobinhoodCrypto @vladtenev
Ad1b.eth | deficloud.eth retweeted
Yo did @priors_trade just stealth launch their own launchpad on the most brutal weekend of the year as a sidequest?
These guys redefine shipping.
robinhood:0xaa07a0e9209e16ac99708c3ec70159c6ef3128a3 robinhood:0xedbf91223639800bcd5756815caf908df3b890be robinhood:0x72be1277e198d0c476869253cb4e62807e14533c snapshot dis
Replying to @miloshquant
Agents will be a big part of how people use onchain finance. Builders are already experimenting with them on Robinhood Chain, and I love seeing it.
5
11
77
Ad1b.eth | deficloud.eth retweeted
It’s AI you stupid bitch, it was always AI, it will forever be AI.
Ad1b.eth | deficloud.eth retweeted
That’s a reversal candle on ETH right there boyos.
Over 2420 we start getting frisky.
Vitalik it’s time for unicorns.
Super is out and the Inu stays Artificial robinhood:0x2e8c31162b855a2ffa90f6f8634643ad6f111e18
1
6
78
Ad1b.eth | deficloud.eth retweeted
Orbio's next chapter starts today.
From a marketplace for AI credit to the front door of private AI.
Every major model in one chat. Turn on Private and your message is encrypted in your browser and opens only inside a verified enclave. Nobody in between can read it. Private runs on open models, uncensored ones included.
It's an agent too. It searches, runs code, ships sites and sends email. Its thinking stays sealed.
First $2 on us, for everyone.
Big update, and we're directly competing with Venice on this one.
Orbio is a privacy OpenRouter with a GPT-like interface.
One chat, with private models at your fingertips: 25 sealed across Phala and NEAR, uncensored ones included, beside every major model below list price.
Your message is encrypted in your browser and opens only inside a verified enclave.
On Venice, by their own description, E2EE needs Pro, and uploads, web search and memory aren't available on it. Here there's no subscription, and the agent keeps working.
Next we're stepping into our own inference for uncensored privacy: distributed GPUs and hardware we run ourselves.
Orbio's next chapter starts today.
From a marketplace for AI credit to the front door of private AI.
Every major model in one chat. Turn on Private and your message is encrypted in your browser and opens only inside a verified enclave. Nobody in between can read it. Private runs on open models, uncensored ones included.
It's an agent too. It searches, runs code, ships sites and sends email. Its thinking stays sealed.
First $2 on us, for everyone.
Ad1b.eth | deficloud.eth retweeted
ai agents have now paid back 400,000 usdg on priors
more than 54,000 loans across 449 agents, and every repayment is public on Robinhood chain
they can put that credit to work hiring each other. an agent lists its work on the board with the link hirers call and what it asks, and its owner can now list it from metamask, rabby or any browser wallet, not only from go mode
the more agents hire each other this way, the more every record says about who to trust
@RobinhoodCrypto @vladtenev @fern $PRIORS
A good read..
If $PRIORS is viewed only as a protocol that provides loans to AI Agents, it is easy to underestimate its potential while also overestimating its maturity too early.
0xedbf91223639800bcd5756815caf908df3b890be
Among the project's recent product updates, I believe the most important development is not simply the gradual improvement of its Agent directory, job marketplace, and payment interfaces, but a set of figures displayed on the official website that can be used to evaluate its business model:
approximately $389,671 in cumulative lending, $387,940 in cumulative repayments, 53,311 recorded loans, and 53,105 recorded repayments. The platform currently lists 315 Agents, 17 backers, and 7 default cases, while reported lender losses stand at $0.
These figures show that Priors has begun facilitating real lending activity. However, whether it can evolve from an early experiment into credit infrastructure for the Agent economy still depends on whether lending activity can continue to grow and whether its credit records can actually be used by other applications.
AI Agents can already call tools, perform tasks, and even pay other services through mechanisms such as x402. However, they still face two practical problems: they need working capital before completing jobs, and they lack reliable credit information when dealing with unfamiliar counterparties. Traditional ratings and reviews are often superficial and easy to manipulate, while actual borrowing and repayment records provide more meaningful evidence of whether an Agent has fulfilled its financial obligations. Priors' core design allows Agents to borrow USDG to pay for APIs, computing power, and tools, repay their loans using income earned after completing tasks, and permanently record their borrowing and repayment histories on-chain. Other applications can query these records to determine whether to extend credit, purchase an Agent's services, or establish new business relationships with it.
I believe the real value of this model is that a single repayment could do more than help an Agent obtain its next loan. It could also improve that Agent's credibility across other applications. If credit histories can only be used within Priors, the project will ultimately remain an independent lending product. But if different applications and lenders are willing to consult the same records, Agents may be able to establish credit identities that work across platforms. Credit would then become more than a statistical record of past behavior. It could become infrastructure that helps Agents access jobs, capital, and trading opportunities.
The recent launch of the Board is an important step in this direction. Priors has begun building a public directory of ERC-8004 Agents on Robinhood Chain, allowing users to examine Agent identities and credit information. For Agents that have borrowed through Priors, the pages can display repayment histories, credit levels, credit progression, and information about the assets and sponsors supporting their credit lines. The project's latest update states that the directory covers nearly 20,000 Agent identities, while earlier disclosures indicated that approximately 302 Agents already had credit records that could be displayed.
These two figures need to be distinguished. Nearly 20,000 registered identities represent potential reach, not 20,000 active borrowers. The approximately 302 Agents with credit histories are a more relevant measure of the group that has begun accumulating actual credit information. The latest website lists 315 Agents, representing another snapshot taken at a different point in time. Figures from different dates and pages need to be interpreted according to their respective definitions rather than combined into a single growth curve. In my view, the actual accumulation of credit records matters more than the size of the directory because it determines how much real data other applications can use when evaluating Agents.
The most valuable aspect of the latest website data is that it allows us to examine whether Priors' lending mechanism is actually operating.
The website reports approximately $389,671 in cumulative loans issued and $387,940 in cumulative repayments, alongside 53,311 loans and 53,105 repayments. Calculated roughly, cumulative repayments equal approximately 99.6% of cumulative lending, leaving a difference of around $1,731. The number of loan records exceeds the number of repayment records by 206.
These ratios look impressive, but I would not directly interpret the 99.6% figure as the final repayment rate. Cumulative lending and cumulative repayments are different aggregate metrics. Outstanding loans, different observation periods, and differences in statistical methodology may all affect the result. The difference between the number of loans and repayments also cannot be directly interpreted as the number of overdue loans or defaults. The website notes that short-term loans are included in these statistics but contribute relatively little to credit scores, further illustrating why a simple comparison of loan counts cannot accurately represent credit quality.
Nevertheless, these figures reveal something worth paying attention to: Priors is no longer just presenting a product interface and a concept for credit scoring. It has accumulated a substantial number of recorded lending and repayment activities. Dividing cumulative lending by the number of loans gives an average loan amount of approximately $7.31 per loan. This looks more like an early-stage market for frequent, small loans than a mature market with significant financing needs. That is not necessarily a bad thing for the Agent economy, because these small loans could correspond to tool calls, API payments, and short-term job expenses. The next question is whether these loans lead to repeat usage and gradually increase in size as Agents take on more valuable work.
The website also lists 17 backers and 7 default cases, while reporting $0 in lender losses. The page explicitly states that these defaults were covered by the collateral staked by the relevant backers. This is a useful data point for evaluating Priors' risk model: it suggests that backer guarantees are not merely a mechanism described in documentation but have already played a role in actual default cases.
However, zero losses for lenders do not mean the system has eliminated risk. The losses have effectively been transferred from lenders to backers, whose staked assets bear the cost. As lending scales, we need to monitor whether backers remain willing to guarantee Agents, whether their collateral is sufficient to cover defaults, and whether concentrated losses could create cascading problems. More importantly, seven default cases are still too few to establish how the model will perform at a much larger scale.
Priors' ten recently submitted ERC proposals are also worth watching. ERC-8458 through ERC-8467 cover Agent income statements, creator fee rights, credit lines, default registries, token launch declarations, liquidity locks, repay-first payees, credit-matched liquidity, Agent credit histories, and Agent credit backers. These proposals attempt to organize the financial activities of Agents into data structures that other developers can understand and adopt. If different lending protocols and applications eventually support compatible standards, an Agent's credit history established on one platform could continue to be useful on another.
However, receiving ERC numbers does not mean these proposals have been formally adopted as industry standards, nor does it mean other projects have already implemented them at scale. The more meaningful catalyst would be external developers and lenders actually writing to and reading these records, allowing Priors' credit system to expand beyond its own product.
Another important development is that Priors is beginning to connect credit with real income from work. The Go Mode job marketplace allows users to post paid tasks for Agents to accept and complete. The project has advertised tasks worth between 1 and 20 USDG, with payments sent to the Agents' registered wallets after completion. Users can also use related tools to discover Agents, pay for tool usage, and confirm borrowing arrangements before transactions take place. Meanwhile, Agents can showcase the services they provide, allowing other users to discover and hire them.
The most interesting part is not that Priors has added a job marketplace, but that work and credit are beginning to form a cycle that can be verified through actual activity. An Agent borrows money to complete a task, the task generates income, that income is used to repay the loan, and the repayment improves its credit history. A stronger credit history may then help the Agent obtain a larger credit line and take on more work. Compared with simply recording a historical score, this mechanism could turn credit from static data into an asset supported by ongoing economic activity.
However, this cycle still needs to be validated with real data. The project previously listed 50 paid tasks, of which five were open at the time. This shows that the product has begun connecting Agents with job demand, but it does not prove that a sufficiently large job market has developed. More important indicators include the number of completed tasks, repeat hiring, the scale of Agents' borrowing, and whether they can consistently repay their loans using income from their work. If the supply of jobs remains limited, even a comprehensive credit history will be difficult to translate into business growth. Conversely, if Agents can consistently secure paid work, demand for Priors' lending services could gradually move from experimental usage toward routine operations.
x402 and MCP are bringing this system closer to real-world workflows. Priors already provides related tools and interfaces that allow Agents to query other Agents' profiles and credit histories, discover paid APIs, and process borrowing, repayment, and service payments. x402 addresses how Agents initiate and complete payments. Priors attempts to address whether they have sufficient funds to begin a task and whether their counterparties have reliable credit histories. Together, these capabilities could support more complex autonomous transactions.
The expansion of collateral options is making the product more concrete as well. Priors already allows Agents to use stock tokens on Robinhood Chain as collateral. The project has given the example that $100 worth of NVIDIA stock tokens could support a credit line of approximately $50, allowing an Agent to obtain working capital without selling its assets. More recently, Pendle's PT-USDG has also been added to the list of eligible collateral. In addition, the project says it checks Chainlink price-feed records over the previous 14 days before issuing loans against stock tokens, attempting to reduce the risk of missing sudden price movements or valuation anomalies.
Expanding the range of collateral means that more Agents and their owners who hold assets may become eligible for credit. It does not mean that lending capacity will necessarily increase at the same rate. Stock tokens remain exposed to market volatility, while PT-USDG involves risks related to maturity, discounts, and liquidity. Valuation and risk-control mechanisms must still be tested through actual lending and repayment activity. Publishing risk-control rules is a positive step, but loan performance, collateral ratios, and default data will ultimately determine whether the model is reliable.
Priors is also continuing to lower its barriers to entry. Google login, gas fees handled by the protocol, a simpler borrowing and repayment process, and automation that only executes after user approval all reduce operational complexity. For infrastructure designed for Agents, these are not minor user-experience improvements. If every loan, payment, and repayment requires extensive manual intervention, Agents will struggle to incorporate Priors into automated workflows. The easier it becomes to integrate the protocol, the greater its chances of moving beyond a small group of developers experimenting with it and toward actual use across more applications.
Returning to PRIORS, the central question remains how product growth translates into token demand. PRIORS is not merely a governance concept. Under the project's design, Agent owners can use the token to support credit lines and potentially obtain larger borrowing limits. According to the proposed mechanism, approximately 75% of the relevant staked tokens may be burned in the event of a default. The token is also connected to planned protocol buybacks, credit backing, and liquidity management.
This design creates a potential link between PRIORS and demand for credit. If more Agents need larger credit lines, their owners may have to commit more tokens as backing. If the default mechanism operates as designed, the token also bears part of the system's risk. Meanwhile, the protocol plans to use a portion of its revenue to purchase PRIORS and incorporate it into its credit-support mechanism. However, collateral demand, protocol buybacks, and liquidity management are distinct mechanisms. They must be assessed separately against the official rules and actual execution records rather than treating planned demand as capital flows that have already occurred.
Revenue allocation also needs to be understood clearly. According to the project's published rules, lending fees are 1% per 30 days. Of this, 50% goes to the providers of lending capital, 25% goes to credit backers, and the remaining 25% goes to the protocol treasury. The treasury could eventually be used for buybacks and liquidity management, but this does not mean that all lending revenue flows directly into PRIORS. Even if the lending business grows, we still need to verify how much revenue actually remains under the protocol's control and how much is converted into token purchases, collateral demand, or reduced supply.
For this reason, I would view PRIORS' value-capture mechanism as a pathway that is still being built, rather than an established flywheel. Product adoption can generate borrowing and job demand, credit lines may drive token collateral requirements, and protocol revenue could support buybacks and liquidity. But each link needs to be supported by actual usage data. If the number of Agents grows without a corresponding increase in genuine borrowing, or if the protocol generates revenue without converting it into observable token demand, product growth will not necessarily translate into a higher token valuation.
There are three catalysts I would pay the most attention to next. First, whether Agents begin borrowing consistently, completing paid tasks, and repaying their loans with income from their work. This will determine whether the credit records are genuinely useful. Second, whether external applications and lenders begin adopting Priors' directory and credit standards. This will determine whether the project can expand from an independent product into shared credit infrastructure. Third, whether collateral demand for PRIORS, protocol revenue, and actual buybacks grow together. This will determine whether the market can shift from pricing expectations for the Agent sector toward evaluating the token based on verifiable economic activity.
The risks are equally clear. Credit records cannot eliminate defaults, and collateral cannot fully protect against price volatility. External applications must actually adopt the proposed standards, the job marketplace needs sustained demand, and early incentives may attract participants who are primarily interested in short-term rewards. Even if the product ultimately succeeds, the token may not capture all of the resulting value. Without transparency around revenue allocation, collateral mechanisms, and buyback execution, it will be difficult for the market to accurately assess PRIORS' long-term economic role.
Overall, I believe PRIORS' fundamentals have strengthened compared with the initial stage, and the improvement goes beyond the number of features released. The public credit directory, job marketplace, MCP and x402 integrations, expanded collateral options, and proposals for shared credit records are gradually connecting lending, income, and credit assessment. The latest website data—53,311 loans, 53,105 repayments, and approximately $389,671 in cumulative lending—also provide evidence that the system is operating in practice. The seven defaults covered by backers offer an initial record of the guarantee mechanism being put to use.
The most important turning point ahead is not how many more features Priors can add, but whether these features can generate sustained real-world transactions. If other applications begin consulting Agents' repayment records on Priors when those Agents work across different platforms, using that information to decide whether to extend credit, purchase services, or establish business relationships, then the value of Priors could extend beyond the size of its own lending business. Conversely, if its credit records remain useful only within the project's own product, while external adoption and real borrowing demand fail to grow, it may ultimately remain an early-stage protocol with a broad feature set but limited network effects.
What makes PRIORS worth betting on, in my view, is the possibility that the Agent economy will eventually need a credit system that different applications can read and use together. The website's figures show that Priors has begun operating, but they have not yet proved that it can scale significantly. The next things to track are whether lending continues to grow, whether income from work can support repayments, whether external applications start relying on its credit records, and whether token collateral and buybacks can develop into verifiable demand as the business expands.
If a single repayment record can eventually help an Agent secure its next job or credit line on another platform, that is when Priors will truly begin evolving from a lending product into credit infrastructure for the Agent economy.
Chart: web3.okx.com/ul/kGC0xUs?ref=…
2
2
45
Ad1b.eth | deficloud.eth retweeted
gm, the next lender your agent meets can read every loan it paid back on priors straight from the chain, and now there's a standard for it 🌿
the erc editors gave the ten drafts we submitted to ethereum/ercs their official numbers!
erc-8458 to erc-8467. each one ships with reference code and tests, and has an open thread on ethereum magicians.
erc-8458 agent income statements
erc-8459 creator fee rights
erc-8460 agent credit lines
erc-8461 agent default registry
erc-8462 token launch declarations
erc-8463 pool liquidity locks
erc-8464 agent repay-first payees
erc-8465 credit-matched liquidity
erc-8466 agent credit history
erc-8467 agent credit backers
these are the parts of priors another lender or app needs, open for any team to build credit for erc-8004 agents the same way. every lender that writes to the same history makes an agent's record count wherever it borrows
302 agents on @RobinhoodCrypto 's robinhood chain already carry that record, with more than 52,000 loans paid back
github.com/ethereum/ercs/pul… $PRIORS
Ad1b.eth | deficloud.eth retweeted
This is independent of Openrouter.
Brought to you by Orbio's Incognito - Private Inference Router
End to end encryption, no third party including Orbio can see your prompts.
Ad1b.eth | deficloud.eth retweeted
Orbio Incognito now works with NEAR AI, beside Phala.
Meet the privacy router: your prompt is encrypted on your device, sealed to the top private models for your mode, and Orbio routes it to one without ever reading it.
auto · cheapest · fastest
Ad1b.eth | deficloud.eth retweeted
The quality legos keep stacking and building off each other.
Next wave on RH will be euphoric, and ceilings much much higher.
We're proud to welcome @orbiodotso as the first launchpad building on Mosh.
Mosh will empower Orbio builders with better market structure & tokenization while the Orbio team empowers them with compute, distribution, and tooling.
robinhood:0xfe7e19cbce2f896c6c528bc355baf5a768291e18
2
4
289
Ad1b.eth | deficloud.eth retweeted
Robinhood Chain will lead the next explosive onchain rally
just like it did the first time around
the same setup is unfolding again, and most people still refuse to believe it
don't say you weren't warned
a lot of you are going to be utterly shocked by just how aggressively explosive the next Robinhood Chain eco pump will be
the kind of move that makes people wish they had bought every dip during the capitulation
this is the accumulation window most will look back on and wish they had taken advantage of
the capitulation is for buying, not panicking
position accordingly

























